For Intelligent and Energy-flexible Production in the Manufacturing Industry

Demand response (DR) programs enable end consumers to adjust their electricity consumption to the power grid’s load and the electricity market’s price signals to reduce their electricity costs. The extension of these programs in Europe is progressing steadily, though slowly.

The aim of the Delfine project, funded by the Federal Ministry for Economic Affairs and Climate Action (BMWK) over three years, was to find technical solutions for industrial customers from various sectors to participate in price-based and incentive-based DR programs. Among other things, the solutions developed make it possible to investigate the effects of DR programs on the balance between electricity generation and consumption as well as on the development of electricity costs, especially in the manufacturing industry.

Positive developments in these areas are a prerequisite for an increased share of renewable energies in the electricity grid. An end-to-end data network was developed using semantic middleware, stretching from the automated creation of production and demand forecasts over the dynamic structuring of electricity prices to the energy-flexible and intelligent use of production resources.

In the project, the FZI focused on predicting load profiles and energy consumption of industrial production processes and, on this basis, a production planning that optimizes energy costs.

At the final meeting, FZI employees Martin Trat and Mischa Ahrens presented the latest machine learning-based prediction and optimization module results. Suitable methods for various use cases were presented and evaluated. In addition, more advanced machine learning approaches, such as transfer learning, were discussed.